香京julia种子在线播放

New and emerging SARS-CoV-2 variants, particularly those with characteristics related to high transmissibility, illness severity, or immune response evasion, have repeatedly threatened the global pandemic response. Thus, the ability to rapidly detect, monitor, and understand the evolution of the virus using WGS has been crucial.

The ability to respond quickly has relied on collaboration between diverse academic and public health professionals to increase sequencing capacity, develop methodologies to further improve the quality and efficiency of the data generated, and perform analyses required for genomic surveillance. Figure 2 shows examples of consortia and networks that have undertaken this work.

The increased sharing of sequencing data and protocols (e.g., https://www.protocols.io) has created an ecosystem of critique and review that has led to continuous improvements to sequencing methods, analytical tools, knowledge sharing, and joint recognition on research publications worldwide. Genomic surveillance during the pandemic was readily achieved globally through the integration of SARS-CoV-2 sequence data into the Global Initiative on Sharing All Influenza Data (GISAID; https://gisaid.org/) database within the World Health Organization (WHO) Global Influenza Surveillance and Response System (GISRS). This initiative was initially established to help understand how influenza viruses evolve and spread during epidemics and pandemics (13). The platform promotes sharing of genetic sequences and clinical and epidemiological data regarding human influenza to link with geographical data associated with avian and animal viruses. GISAID involves public-private partnerships and academic institutions in many countries supporting data sharing and collaboration among researchers. A GISAID SARS-CoV-2 database was quickly implemented based on the influenza model; it was extremely useful, but criticisms were raised about governance and a lack of full public availability of data.

The sharing of SARS-CoV-2 sequence data via the GISAID database enabled countries around the world to understand which variants were emerging and circulating nationally and internationally. This allowed for early responses to mitigate the transmission of novel variants across borders, and for the amendment of testing and sequencing procedures to detect new variants. This was exemplified by the identification of the Alpha variant using WGS data generated by the COVID-19 Genomics UK (COG-UK) Consortium, which explained changes to the epidemiological characteristics and transmission dynamics of SARS-CoV-2, as well as the later identification and tracking of the Delta and Omicron variants of concern (VOCs) (14). Analysis of SARS-CoV-2 genomes and linkage to clinical data have informed medical improvements, including evaluations of vaccine efficacy against specific variants of SARS-CoV-2 and research on the susceptibility of viral variants to therapeutics against COVID-19 (8, 15–18).

Pre-existing sequencing infrastructure and expertise in pathogen genomics and bioinformatics within academic institutions and public health institutes were key success drivers for countries that rapidly introduced genomic surveillance of SARS-CoV-2. There was an urgent need to handle, store, and interpret large genomics datasets to enable genome-based decision-making processes. In the United Kingdom, the Cloud Institute for Microbial Bioinformatics was launched in 2014, with support from the Medical Research Council, and this pre-existing open, cloud-based infrastructure and expertise enabled COG-UK to rapidly share genome data and develop tools to interpret large datasets in a secure environment. In addition, the ARTIC network (https://artic.network/), built on previous experience in responding to other viral outbreaks, rapidly developed sequencing protocols and analytic tools for SARS-CoV-2 that were used worldwide. This highlights the importance of continued investments to support technical infrastructure, methodological development, training, and public health decision-making for genomic surveillance (19). Such investments reduce sequencing costs and turnaround times, provide better flexibility in sample types, and improve data quality, thereby increasing the affordability, scalability, and robustness of sequencing procedures (20, 21).

As shown in the United Kingdom and Africa, distributed sequencing through a “hub and spoke” model enabled regional sequencing sites to ensure technical capabilities and expertise, and to work closely with local hospitals to rapidly sequence and analyze WGS data, while enabling public health agencies to focus on national public health decision-making (19, 22). The relationships fostered between healthcare and academic professionals demonstrate the value of closer intersectoral links and the alignment of clinical care with research and innovation.

Public health services faced a complex and sometimes conflicting array of priorities during the COVID-19 pandemic, including the testing and sequencing of samples from populations in the community, in healthcare settings, and at borders. This resulted in governance and management challenges that hindered decision-making and the streamlining of processes.

Key vulnerabilities in genomic surveillance systems were recognized during the COVID-19 pandemic. Access to reagents and consumables necessary for sequencing was hindered by global shortages, limiting sequencing capacity and revealing a need to strengthen supply chain resilience akin to corresponding reflections regarding diagnostics and pharmaceuticals (23–25). Better communication and collaboration between laboratories, as well as a central and more sophisticated stock management system, could enable better resource management across laboratories (22, 26). The lack of integration of surveillance systems and healthcare systems at a national level, together with a lack of data sharing standards, hampered data sharing and the linkage of clinical and sequencing data in many countries at the beginning of the pandemic. Further advances in informatics and interoperability between databases are needed to link large genomic datasets with other research and clinical datasets to support clinical decision-making and multi-layered research investigations. Challenges were also encountered with sample collection, sample storage, and logistics. These included the identification of a suitable point of contact at each testing laboratory, data protection and confidentiality issues, ethical clearances to access and send samples with corresponding metadata, access to affordable transport that could provide dry ice when required, and point-to-point delivery of samples. COVID-19-related staff absences further stretched workforce capacity, leading to staff feeling overwhelmed by requests to transfer samples with corresponding clinical metadata to national surveillance systems (27), added on to routine clinical testing. Owing to the lack of guidance for reporting WGS results from academic institutions, reports were often not standardized and there was no clear process for validating them (21, 26).

The worldwide deployment of SARS-CoV-2 genomic surveillance capacities was geographically uneven and followed a variety of implementation strategies adapted to national health system organizations and resources. Available sequencing capacities could not readily meet the increasing demands owing to limitations in infrastructure and trained personnel. This proved particularly problematic in low-income settings, for example in many parts of Africa, where a lack of existing sequencing infrastructure and expertise in genome analyses hindered the management and initial response to COVID-19 (22). Global analysis of viral sequence data deposited in public repositories has allowed quantitative assessments of national SARS-CoV-2 genomic surveillance programs in terms of case coverage, timeliness of reporting, and completeness of data sharing, broken down by sampling strategies and sequencing technology (28–30). The analysis has shown that, during the first 2 years of the pandemic, the SARS-CoV-2 genomic surveillance landscape displayed marked heterogeneity of outputs by country and economic income level. While 78% of high-income countries (HICs) sequenced >0.5% of reported COVID-19 cases, only 42% of low- and middle-income countries (LMICs) achieved that coverage level (29). The mean sequencing coverage ranged by region, from 0.1% in the Eastern Mediterranean region to 3.4% in Europe (28). Mean turnaround reporting time was shorter in HICs (24 days) and in Europe (18 days) than in other income groups or regions (28). Coverage increased and turnaround times shortened gradually during the pandemic (29, 30).

Global SARS-CoV-2 data sharing was also very uneven geographically. Uploading of VOC sequences in the public domain was incomplete in many countries and metadata release into GISAID was inadequate, with 63% of global sequences missing patient demographic information and more than 95% missing clinical information (28).

In Africa during 2020, 72% of genome sequencing capacity was concentrated in four countries (31). The SARS-CoV-2 genomic surveillance capacity rapidly expanded after coordination and mobilization of resources from 55 countries, with logistical support by the Africa Centres for Disease Control and Prevention (CDC), the WHO, and public-private funding, as part of the African Pathogen Genomics Initiative (31–34). This was achieved through increased continental coordination, strengthening of infrastructure for sequencing and data sharing, and workforce training in genomics and bioinformatics. Turnaround times sufficient to inform public health decisions were related to proximal access to sequencing services (32). The African SARS-CoV-2 genomic surveillance collaboration delivered important early warnings by discovering several emerging VOCs in South Africa (32).

Common implementation enablers can be learned from these experiences. In high-, low-, and middle-income countries, mobilizing sequencing expertise and resources across sectors and jurisdictions, adopting representative and targeted national sampling strategies, investing in sequencing infrastructure, analytics, and expert workforce training, and providing public health surveillance centers access to data sharing and analysis platforms were key improvements that allowed the gradual implementation of effective national and regional SARS-CoV-2 genomic surveillance programs for early detection and tracking of new variants of interest (16, 19–21, 26, 32, 35–37).

Current WGS methods have revolutionized the design and efficiency of pathogen surveillance by accelerating the sequencing and analysis process and increasing sample throughput. Samples are no longer collected by clinical and epidemiological suspicion: rather, prospective isolate collections are massively genome-sequenced and subjected to automated and parallelized data analysis to investigate the phylogenetic relatedness of strains, generate transmission hypotheses, and identify potential sources. This has been successfully introduced for genome-based surveillance of endemic diseases and conditions, e.g., foodborne and sexually transmissible pathogens, Mycobacterium tuberculosis, and AMR. WGS is also a key tool for the investigation and public health management of emerging pathogens, as was illustrated during the recent mpox global epidemic (49). WGS methods first used a decade ago in a limited number of well-equipped public health laboratories in high-income countries are now being successfully deployed in most regions of the world (50–54). Genome-wide high-resolution pathogen typing and surveillance has been implemented in many national and international action plans and is supported by leading health organizations such as the WHO (55–57), the United States CDC (58), the European Centre for Disease Prevention and Control (ECDC) (10, 59), and the United Kingdom National Institute for Health and Care Research (60). This new strategic approach foresees an epidemiological confirmation of genome-based transmission hypotheses and transforms pathogen surveillance from monitoring disease incidence trends to delineating their transmission pathways at a highly scalable resolution level (9, 42, 43).

Pathogen genomic surveys and surveillance programs have shown the value of genomic analyses to monitor pathogen diversity and evolution, track population distribution of lineages and phenotypic variants of public health importance, characterize changes in virulence factors, and predict antigen matches with current and candidate vaccines and neutralizing antibodies, both for viral and bacterial diseases. The integration between epidemiology and phylogenetic analysis into the evolutionary epidemiology approach, using phylogeographic and phylodynamic computational analyses, has been widely applied to decipher the evolution of genomic diversity and reconstruct routes and timing of transmission of major viral pathogens, including the influenza virus, hepatitis viruses, human immunodeficiency virus, Zika, Chikungunya (61), SARS-CoV-2 (62–66), and mpox viruses (49, 67). Genomic surveillance of the influenza virus has become an integral part of GISRS to inform the yearly vaccine design cycle, by continuous monitoring of the rate of viral evolution and identifying novel variants, for both influenza and SARS-CoV-2 viruses (56). Phylogenetic and phylodynamic studies of SARS-CoV-2 virus diversification, lineage transmissibility, and fitness can help design vaccines, monitor the cross-immunity of circulating lineages, and evaluate vaccine selective pressure on lineage-specific, immune escape mutations that may affect the efficacy of vaccines and therapeutic monoclonal antibodies (5, 7, 15, 18, 68). For example, the identification of mutations in the spike protein of SARS CoV-2 has informed our understanding of immunity in populations and the adaptation of diagnostic assays and vaccine production (69–72). Further integrating high-resolution pathogen phylogenomic data into mathematical disease transmission models offers potential for more precise forecasting and prediction of infectious disease epidemics and the effect of health interventions (73).

The era of vaccine development based on molecular and genome data was introduced almost 20 years ago under the term “reverse vaccinology” (74). The use of genome data again pushed this field forward (75), leading to innovations such as genome-guided RNA vaccines against cancers (76, 77). Prominent examples of gene-guided vaccine development already exist for bacterial pathogens. Genomic surveillance is increasingly informing both vaccine development and adaptive updates of antigen composition and vaccination schemes to improve and sustain prevention of bacterial diseases. WGS analysis has become of paramount importance in monitoring vaccine-preventable bacterial pathogen diversity and population dynamics, and in identifying virulence traits and determinants of immune escape, including antigen gene mutations, recombination, and deletion in major pathogens, such as Streptococcus pneumoniae (78), Neisseria meningitidis (79, 80), and Bordetella pertussis (81, 82).

By rapidly and inexpensively characterizing the entire genome of a pathogen, WGS data analyses provide attractive tools for comprehensive, fast, and reliable microbial diagnostics (46, 49). The genome-wide approach has allowed the simultaneous prediction of drug resistance to guide antimicrobial therapy and virulence assessment (e.g., the production of toxins) to support adjunctive patient treatment (83–85). Reasonable attempts have been undertaken to derive the huge armamentarium of possible resistance mutations and genes from genome data. Web-based tools and international databases that predict antimicrobial susceptibilities are continuously updated, assessed, and validated and have reached a level of accuracy comparable or even superior to phenotypic susceptibility testing (86–88). Even multifaceted, complex, and multifactorial resistance mechanisms can be predicted from microbial genome data as an input (89–91). High-performance computing using machine learning algorithms and artificial intelligence (AI) allows AMR to be derived from large genome datasets, even in the absence of knowledge on resistance mechanisms or reference templates (92–95).

WGS data can also guide the development of novel diagnostics. Analysis of the genomic information of an emerging pathogen (or a new variant thereof) in comparison with a large dataset of already available genome data allows the derivation of discriminatory and diagnostically relevant markers (11, 96, 97).

Classical diagnostic procedures, some also using WGS data, mainly started with the isolated and identified strain of a microorganism. However, WGS-based approaches do not necessarily require a pure culture of an isolate ( Figure 3 ). Significant advancements have been made in the last 10 years to improve culture-independent diagnostics and strain typing. Finding pathogen-specific signatures in complex sample material paved the way for metagenomic diagnostics and strain typing to guide therapy and infection prevention and control measures in the veterinary, food safety, and public health sectors (39, 46, 62, 96, 98). Meanwhile, metagenomic diagnostics is becoming an established field in human medical and veterinary microbiology with an immense potential that is actively advanced by an engaged scientific community. Metagenomic analysis of human, animal, and environmental samples also offers high sensitivity and flexibility for detection and quantitation of known and novel pathogens for public health surveillance (99). Requirements for standardization (in sampling, data generation, and analyses), comparatively high analytical cost, and certain legal and regulatory barriers currently hinder a wider medical application of this approach (100, 101). However, these barriers are expected to be resolved in a stepwise manner in the coming years, building on current standardization steps (102).

Technically, next-generation sequencing (NGS) describes a procedure involving the massive and parallel amplification and subsequent sequencing of DNA fragments of varying but comparably small lengths. This is accompanied by some limitations and drawbacks for downstream processing (103), for example, assembly breaks at repetitive elements and a missing differentiation between the chromosomal and plasmid contigs. Long-read sequencing (third-generation sequencing) circumvents these problems, but at the expense of higher quality input samples, higher costs, and higher computational demands, as well as potential for lower sequencing accuracy. Newer long-read sequencing technologies, such as those developed by Oxford Nanopore Technologies (ONT), have now reduced costs and improved accuracy, scalability, affordability, and flexibility (104). They are readily employable in low-resource settings (46). Until now, long-read sequencing projects have been mainly research-oriented but, with the possibilities of multiplexing and parallelization, the cost-per-sample can be drastically reduced and throughput greatly increased (105, 106). Combining the analysis of long- and short-read data in an aggregated form has been a useful approach for diagnostic and public health-oriented WGS applications, combining the advantages of both technologies for outbreak and phylogenetic analyses, resistance and virulence prediction, and genomic pathogen surveillance (107–109). Long-read sequencing using nanopore-based technologies with software optimization reaches high accuracy levels that may equal those of short-read sequencing data, removing the need for the latter for applications such as determination of bacterial drug susceptibility and delineation of transmission clusters in outbreak investigations (104). Recent technological advancements such as ultra-rapid genomic sequencing, single-cell sequencing, and sophisticated AI/machine learning algorithms for reference-free and stringent data analysis are propelling this field forward (110–114).

The utility of pathogen sequencing data significantly increases when it is linked to clinical, phenotypic, or epidemiological data (i.e., “metadata”). Such data linkage between healthcare records in agencies and institutions can improve health outcomes and the effectiveness of public health surveillance—delivering clinical, public health, and economic benefits (115–118). For example, genomic data sharing has illuminated multi-jurisdictional outbreaks and cross-border spread of communicable diseases (11, 63, 119, 120).

Figure 4 outlines the integration of laboratory, clinical, and epidemiological metadata with pathogen sequencing data obtained from clinical samples using different sequencing approaches. The sequencing data include files generated by sequencing instruments (i.e., “raw” sequencing data stored as fastQ files) and inferred by bioinformatics pipelines (e.g., consensus genomes stored as fastA files, variant files, or network representation files). The genomic data are accompanied by laboratory findings obtained from other frontline diagnostic tests on the sample and stored in the laboratory information systems. For example, the genomic sequence of M. tuberculosis should be linked to results of associated phenotypic antimicrobial susceptibility testing (83). Other types of metadata include case demographics, clinical diagnoses with co-morbidities coded using international standards, past exposures to infected cases, treatment outcomes, and use of healthcare resources (e.g., hospital admissions and/or vaccinations). There is also a case for integrating disease occurrence metadata into relevant and validated phylogenetic models to improve epidemiological inferences from pathogen genomes (65). Occurrence metadata reflect background prevalence of the infection or disease and capture cases diagnosed by frontline diagnostic tests but without accompanying sequenced genomes and the time of their diagnosis.

The amount of microbial sequencing data generated in clinical, public health, and research laboratories has been rapidly expanding. Even before the COVID-19 pandemic, it was suggested that genomics would produce more data than social media, earth sciences, or astronomy (121). These projections have been fulfilled by the immense volumes gathered by the ongoing genomic surveillance of SARS-CoV-2 variants. Open databases such as GISAID now contain over 16 million SARS-CoV-2 genomes recovered from cases across all continents.

Data integration requires digital infrastructure for the collection, storage, and linkage of sequencing data and sample-associated metadata, as well as the sentinel and syndromic disease surveillance data produced across national and international networks (122, 123). As the data are produced by multiple pathology and clinical service providers, federated data storage solutions are usually employed and optimized for data heterogeneity and volume (e.g., genomic data warehouses). The portability of de-identified and often encrypted health data uploaded to the cloud can provide significant leverage to a growing number of data users interested in secondary analyses of the pathogen sequencing data for infectious disease management, prevention, and research. These analyses demand large-scale data aggregation and the use of third-party algorithms and tools designed for users with different backgrounds (39, 64).

Integration and responsible sharing of sequencing data should be governed by the findability, accessibility, interoperability, and reusability (FAIR) principles of Open Science. The core elements of such data sharing include transparency regarding the purpose, collection, use, and exchange of data; data quality, standardization, and security; and recognition and attribution, thereby fostering trust, integrity, and reciprocity. The sharing and mobilization of data empowers research and development, but genomic data generated and shared by data donors must also comply with relevant local policies and regulations.

The uptake and success of national genomic surveillance systems depends on the access to sequencing and data sharing across different sectors and jurisdictions. The main success factors are metadata and sample volume, density of sequencing, and timeliness of sequencing and data sharing.

The more contextual metadata are linked to genomic sequencing data, the greater the accuracy of a surveillance signal (124) and the potential value of such combined data for decision-making (125). The value of sequencing data is also directly related to the volume of shared data available for analysis. Larger sets of shared genome sequences are more representative of current disease activity and are more likely to allow the identification of clusters of infections with a common source. For example, for every 1000 genomes of foodborne bacteria (e.g., Listeria monocytogenes) added to a database, there is a reduction of approximately six cases of illness per pathogen per year (126). The utility of pathogen sequencing data is also a function of sequencing density (i.e., the proportion of cases of a disease from which genomes of relevant pathogens are sequenced and shared) and the timeliness of data sharing ( Figure 5 ). The shorter the time between outbreak onset and the sharing of pathogen sequences associated with the outbreak, the higher the utility of these sequences for infection control and public health response. Density of sampling reflects the local or national ability to sequence target genomes in positive clinical samples from a representative proportion of diagnosed cases. The sequencing density varies for different diseases and can be limited by access to positive clinical samples and genome-sequencing technology.

New applications continue to emerge to exploit the ultimate resolution offered by WGS, such as the detection of emerging pathogen variants or mixed infections (127). Nevertheless, the design of genomic surveillance systems should be guided by required sensitivity, specificity, and timeliness, as well as by national capacity in pathogen sequencing, to match local public health response capacity and to address specific operational public health objectives. The detection of clusters in an outbreak can be delayed if sampling density is limited and few cases are sequenced. The resolution of outbreak detection necessary to support public health responses can be estimated using both real and simulated genomic data. This approach informed the design of SARS-CoV-2 genomic surveillance sampling strategies in countries with adequate SARS-CoV-2 polymerase chain reaction (PCR) testing and sequencing capacity: simulations demonstrated that the impact of sequencing fewer cases depends on the size of the outbreaks, and on the genetic and spatiotemporal similarity of index cases of the outbreaks (128). Furthermore, these assessments can quantify the minimum number of sequenced cases required to detect and monitor new virus variants, while also considering representative sampling of cases and potential sampling biases. Insufficient timeliness of genome sequencing and reporting was found to be a frequent limiting factor for practical use of SARS-CoV-2 genomic data to inform transmission control interventions in acute healthcare settings (129).

The design of laboratory surveillance systems should balance national needs of data synthesis and response coordination with jurisdictional priorities for informing local public health responses and managing organizational data assets (57, 130). In this context, genomic data and metadata sharing often revolves around existing health data protection regulations. While some successful national systems are very centralized, others are based on decentralized primary bioinformatics analyses, open-source cloud-hosted shared pipelines, and restricted exchange of sequence and metadata between federal member states as well as across sectors—with shrinking data exchange volume from local to federal levels (131).

Benefits of genomic data sharing include more efficient public health surveillance (enabling better targeted and nuanced interventions or “precision public health”), the ability to examine local data in the global context in real time, and opportunities for research and development from data re-use (4, 12, 132, 133). The risks of data sharing are largely of concern to data donors and relate to patient privacy, confidentiality, and data protection and security. Although the risk of re-identification of patients from publicly shared health data is considered low, the risk assessment does not merely depend on the statistical probability of record attribution; it is also affected by the sensitivity of the data. For example, for COVID-19—a condition with a relatively small degree of potential invasion of privacy—the risk of re-identification of <10% has been considered acceptable. For high sensitivity data (e.g., associated with sexually transmitted infections) some guidelines recommend a threshold of 5% (134). Re-identification risks can also be reduced by the removal of human-like DNA sequences from the microbial sequence files before archiving them in any public database. Sequencing data donors have also been increasingly concerned about legal and ethical implications of incidental findings by data recipients, misuse of data, profiteering from re-use of data with limited attribution of intellectual property, and a lack of accountability among data recipients owing to distributed and unpredictable data re-use (135, 136).

National agencies are responsible for collating and mobilizing public health data to improve population health, as data hoarding and poor data stewardship can hinder genomic surveillance (137, 138). However, the collation and storage of extensive metadata for individual genomic results can be laborious and escalates the consequences of data security breaches. To minimize such risks and costs, only minimal sets of metadata are usually shared with microbial genomes, especially in publicly accessible databases. For example, for SARS-CoV-2 genomic surveillance, minimum contextual data could consist of the patient’s age (often displayed in 10-year intervals), gender, sample collection date, geographical region of collection, and the variant/lineage designation of the microorganism. Such sharing of microbial genomic data and metadata associated with diseases affecting humans and animals, and involving cases across different jurisdictions, requires trust and collaboration between all relevant sectors (i.e., human and animal health, food, and environment) and stakeholders (e.g., government, commercial, and not-for-profit entities) (139).

Consensus-based guidelines are needed to ensure responsible integration and mobilization of pathogen sequencing data, in accordance with ethical and legal requirements that guarantee equal and fair distribution of risks and benefits without an increase in global inequality. The anticipatory use of integrated data analysis for comprehensive disease surveillance and awareness will facilitate better informed and more efficient mitigation of the risk and impact of future pandemics.

“One Health” is an integrated, unifying approach that aims to balance and optimize the health of people, animals, and the environment (140). Both “One Health” and “Planetary Health” concepts have gained prominence as we acknowledge how people, animals, plants, and our environment are interconnected and interact, and how changes to ecosystems can create emerging threats with serious consequences for our health. The COVID-19 pandemic exemplified the importance of using a One Health approach to prevent, predict, detect, and respond to emerging global zoonotic health threats. SARS-CoV-2 is an example of a zoonotic pathogen (2) and human–animal and animal–human transmission events were detected during epidemic waves, e.g., in Denmark (141). Indeed, the majority of emerging pathogens are zoonotic and arise or originate from an animal source or reservoir. WGS has shown diagnostic value for early warning of potential cross-species spillover events of zoonotic pathogens, such as avian influenza (142, 143). A One Health approach is vital for ensuring infectious disease preparedness. To achieve this goal, developing and deploying “agnostic” or broad-range targeted metagenomic sequencing of samples from atypical diseases in humans and animals as well as from their environment has been advocated as the most promising One Health preparedness strategy for detecting the emergence of novel Disease X (99, 144).

AMR is also an archetypical One Health issue by nature. A recent, mixed-method study estimated that 4.95 million (95% uncertainty interval 3.62–6.57) deaths were associated with bacterial AMR in 2019 globally, including 1.27 million (0.911–1.71) deaths attributable to bacterial AMR (3). Excessive use of antibiotics in healthcare, animal husbandry, and aquaculture all contribute to the increasing levels of AMR observed worldwide in many pathogens. Indeed, globally, two thirds of antibiotic usage is agricultural (i.e., for farm animals and crops) (145). Moreover, resistant organisms and genetic determinants of AMR spread across ecosystems—i.e., between animal and environmental reservoirs and human populations via various pathways (146). A One Health approach to AMR surveillance and control is therefore vital ( Figure 6 ). It requires an improved understanding of these pathways and how they can be tackled. It is also crucial to understand how environmental pressure gives rise to new resistance mechanisms.

The One Health approach mobilizes multiple sectors, disciplines, and communities at various levels of society to work together to understand and address the root causes of disease and generate long-term sustainable solutions. Conceptually, a global One Health collaboration has been formalized by WHO, the Food and Agriculture Organization of the United Nations (FAO), the United Nations Environment Programme (UNEP), and the World Organization for Animal Health (WOAH) (147). This “One Health Quadripartite” aims to support multisectoral approaches to reduce health threats at the human-animal-ecosystem interface, coordinate and engage partnerships, and facilitate interactions of the multiple objectives with a central goal. Moreover, the regional centers for disease control in many WHO regions have already adopted One Health programs for disease surveillance, e.g., the United States CDC (148), ECDC (149), and Africa CDC (150). In Europe, European Union (EU) agencies across relevant sectors have established a One Health cross-agency task force (151, 152). At the country level, strengthening the intersectoral collaboration and communication between different sectors and actors is also paramount. Countries are encouraged to work together on designing and improving their One Health approaches by sharing experiences and best practices.

Figure 6 illustrates how One Health genomic surveillance needs to span across sectors. Several separate surveillance schemes for human and animal pathogens are well established. In contrast, the sharing and usage of surveillance data across the sectors is less well explored, even though various models exist to harness high-throughput sequencing technology for integrated surveillance. FAIR principles can guide improvements in the sharing and usage of data across sectors.

Pathogens and AMR determinants can often be directly detected using the same molecular methods regardless of whether the sample is from a human, an animal, or the environment. This is particularly relevant to metagenomic pathogen surveillance (99, 153). This applies to methods used in bioinformatic pipelines and phylogenetic analyses. However, sample collection, preparation, and the interpretation of results may require specific approaches and knowledge depending on the surveillance objective. Harmonizing methods requires—and supports—cross-sectorial collaborations. Application of the same methods in laboratories across sectors and for different sample types, as well as the establishment of One Health laboratories, supports the adaptability of these systems to new pathogens and may provide opportunities for flexibly redirecting or scaling up capacities.

Harmonizing surveillance methods across sectors enables the generation of comparable and interoperable data from different sources. Nevertheless, sharing and using data for surveillance across sectors often requires an understanding of relevant intersectoral differences, for example regarding biological characteristics of hosts and pathogens, sampling strategies, and terminology (154).

In the United States, real-time genomic surveillance programs for foodborne pathogens have been developed in the last decade to support food safety and public health, including the GenomeTrakr program operated by the Food and Drug Administration (FDA) and the PulseNet program coordinated by the CDC (126, 155). Similar national programs for One Health real-time genomic surveillance of foodborne infections have also been successfully developed in other areas, including Europe and Australia (50, 119). These programs have demonstrated enhanced sensitivity over traditional surveillance for early outbreak detection and have proved effective, e.g., in preventing foodborne infections caused by Salmonella (119, 156), L. monocytogenes (157), and Shiga toxin-producing Escherichia coli (STEC) (158). They also showed cost-effectiveness when optimized WGS technology was used appropriately (159).

Wastewater surveillance is a timely example of an approach that can be a cost-effective, complementary component of integrated One Health surveillance of infectious diseases and AMR. Wastewater molecular analysis by PCR and metagenomic sequencing can provide a variant-specific, community-representative picture of human pathogen prevalence and early warning of local spread, as shown for SARS-CoV-2, influenza, mpox, and vaccine-derived polio viruses (153). Wastewater surveillance for mapping SARS-CoV-2 distribution and the magnitude of the infected population exemplifies how rapidly changing pathogens can be tracked across large populations (160). Wastewater-based surveillance has been shown to be useful at time points with less extensive testing, providing a quick overview of the geographical spread of, e.g., a new SARS-CoV-2 variant (8). After having been widely applied for the surveillance of SARS-CoV-2 in numerous countries, it is time to focus on harmonizing the approaches and widen the use of the approach to other pathogens, including emerging new pathogens and variants (153). In Europe, a large consortium (EU-WISH) with this aim was launched in 2023.

Wastewater surveillance has also been shown to be a promising tool for broad range monitoring of bacterial AMR. Using metagenomic analysis of AMR gene distribution in urban sewage around the world, Hendriksen et al. observed that local sanitation standards and healthcare conditions correlate with the abundance and type of AMR determinants (90). This mapping of local resistome profiles has expanded to over 100 countries (161). Such a metagenomic epidemiological approach measures the collective signature of microbial carriage across a community. However, many aspects are still unresolved. Current technology is not clear in terms of sensitivity, specificity, and plausibility. Since AMR pathogens may also propagate and live as saprophytic bacteria in the aquatic environment and sewage, wastewater surveillance of AMR fundamentally differs from that of viral pathogens. Therefore, there are still challenges in developing standardized and optimized methodologies for wastewater AMR metagenomic surveillance and in interpretation of data for conducting quantitative microbial risk assessment.

As discussed above, international genomic surveillance during the COVID-19 pandemic was rapidly achieved through the release of SARS-CoV-2 data into the GISAID database. Additional tools enabling real-time genomic surveillance of COVID-19 at a global level include the public availability of sequence-based consensus nomenclature, such as Pango Lineages databases (https://cov-lineages.org) (68) for labeling and open access analysis, and visualization tools (e.g., www.nextstrain.org) for tracking viral phylogenetic lineages and variants. Consensus nomenclature was established by WHO with a multidisciplinary panel of experts to define and label SARS-CoV-2 genomic variants of interest and variants of concern, based on increased virulence, transmissibility, or immune escape (https://www.who.int/en/activities/tracking-SARS-CoV-2-variants/). Similar sets of publicly available databases and scientific resources are available for global genomic surveillance of other pathogens, including microbial genome databases such as PubMLST, which provides a gene-by-gene analysis-based strain nomenclature (173). Interactive online platforms and bioinformatics toolboxes such as Pathogenwatch (https://pathogen.watch/) and the Center for Genomic Epidemiology (https://www.genomicepidemiology.org) have been developed for integration and visualization of bacterial genomic and epidemiological data, and for the identification of high-risk epidemic clones and lineages (174).

We propose that the worldwide development of routine genomic surveillance for SARS-CoV-2 needs to be expanded to other epidemic-prone pathogens, for the tracking of AMR mechanisms, and for future pandemic preparedness. Clearly, this effort would require wider global public health and One Health commitment and agreement on technical standards for the responsible, voluntary, and secure sharing of coherent genome and health data beyond SARS-CoV-2 (175). Demonstrating the positive health impact and cost-effectiveness of local, national, and international genomic surveillance for different pathogens and AMR threats will help guide what data need to be collected and shared in a timely manner.

Recently, WHO, in line with the World Health Assembly Resolution 74.7, enacted governance principles for international genomic surveillance data sharing as part of a 10-year strategy for global pathogen genomic surveillance for epidemic preparedness (130). Novel resource centers have been established recently to support implementation processes, e.g., the WHO Hub for Pandemic and Epidemic Intelligence and the WHO BioHub System (https://pandemichub.who.int/pandemic-hub) for international sharing of biological samples (https://www.who.int/initiatives/who-biohub). Academic, public health, and non-governmental organizations should work in sharing best practices and technical solutions toward supporting the equitable deployment of pathogen genomics capabilities through collaborative initiatives at the regional and global levels, including well-established regional public health surveillance networks (167, 176) and the recently established International Pathogen Surveillance Network (IPSN) (177).

WHO has long-established flexible AMR phenotypic data reporting tools (e.g., WHONET 2022) for the Global Antimicrobial Resistance and Use Surveillance System (GLASS). They published an expert opinion on potential approaches to further integrate reporting of WGS data on AMR determinants into GLASS (55). A number of WHO Collaborating Centres for Antimicrobial Resistance Epidemiology and Surveillance (https://www.who.int/about/collaboration/collaborating-centres) support the technical implementation of the WHO Global strategic action plan on AMR and GLASS national programs.

Key efforts to build new public health infrastructures and processes for using actionable outputs from genomic surveillance should focus on enhancing the understanding of clinical and public health practitioners and health officials across sectors on how to interpret genomic epidemiology findings and translate them into action. First, this should include the development of novel training curricula and programs for surveillance data end-user health professionals as well as advanced meta-disciplinary specialty training for the new generation of genomic surveillance experts, such as genomic epidemiologists and public health bioinformaticians. Second, efforts should promote further collaboration and dialogue among virologists, microbiologists, evolutionary biologists, environmental scientists, epidemiologists, and public health and animal health practitioners into a transdisciplinary One Health community of practice. Third, connections should be strengthened between surveillance networks and open data research platforms hosting curated databases of genomic sequence data and interactive bioinformatic nomenclature assignment resources for specific viral, bacterial, fungal, and parasitic pathogens of public health importance. Fourth, external quality assessment and interlaboratory ring tests should be implemented to ensure reproducibility and comparability of derived consensus sequence data and assembly pipelines. It is key that these multilateral and multidisciplinary collaboration efforts build on the existing disease-specific research and surveillance networks, and that they benefit from close interaction with the cognate scientific as well as professional healthcare community (10, 57, 59).

Further international collaboration is needed on developing common methods and nomenclature for genomic surveillance of epidemic-prone infectious diseases and multidrug-resistant organisms of public health concern. These efforts encompass the harmonization of epidemiological and environmental sampling schemes for surveillance programs, as well as transparency of bioinformatic analysis pipelines and agreement on microbiological, clinical, and epidemiological data semantics.

Gaps in genomic data management, sharing, and analytic capabilities are common barriers to wider implementation of real-time pathogen genomic surveillance of SARS-CoV-2 and other pathogens, as noted above. These gaps underline the added value of further developing unified and freely accessible bio-informatics ecosystems for public health agencies, first available at national and regional levels, that provide open-source integrated data processing, data sharing, genomic analytics, and reporting tools, including standardized metadata and automation support by application programming interfaces (APIs) (10, 183).

To enhance interpretation of pathogen and AMR genomic surveillance data and their translation into public health risk assessment and mitigation, it is also desirable to facilitate rapid data deposition in public sequence archives, such as the United States National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/) and the European Nucleotide Archive (https://www.ebi.ac.uk/). These repositories, in turn, enable data use for research with scientifically curated genomic data platforms and genetic nomenclature resources. This allows inferences to be drawn on pathogen evolution, functional prediction of AMR, virulence, and antigen phenotype and the tracking of lineages causing disease across countries and settings. Currently, genomic sequence data can be directly deposited there using private subsets functionalities. Examples of open science resources include projects such as the K. pneumoniae Global Genomic Surveillance Platform (https://klebnet.org/) and scientific discussion forums, such as the Virological forum (https://virological.org/about), which allow for open analysis and interpretation of virus molecular evolution and epidemiology.